Search sheets for group text generation

Ben Swift · 6 August 2026

The search sheets are the cutouts activity with the cutting taken out, for groups to use together. The CLI shuffles a corpus and deals it round-robin into one page per participant, so no two sheets are the same. You call out the last token written on the board, everyone scans their own sheet for it, and hands go up wherever there’s a match. Pick one of those hands at random and you have your next token.

Every token pair in the text is printed on exactly one sheet. The CLI spreads each context across as many sheets as possible, so common continuations tend to get more hands than rare ones. Picking a hand at random approximates sampling from the model’s probability distribution. Take whoever shouts first and you’re sampling the fastest reader instead of the text.

Each participant holds part of the model on their sheet. If someone is absent, the group loses access to the token pairs on that page.

Four sets are ready to print: Green Eggs and Ham at 15 sheets, Peter Rabbit at 17, The Cat in the Hat at 36 (A5 sheets two-up on A4), and Australia at 79. Print them single-sided and hand out one each. The first page is the briefing for whoever is running it, and every page after that is a different participant’s sheet. Every sheet carries the same number of token pairs, so the length of the text decides how many sheets there are, and a set only suits a group of about that size. The tools page will build a set from any text you paste in.

Building these turned up a problem with the generated colour palette. The sheets use 16pt type for token pairs, compared with 36pt on the cutouts. At that size, I struggled to distinguish colours that were well separated by the OKLab ΔE metric. The palette included several shades of the same hue, which were hard to tell apart when scanning small text.

I changed the search to start with colour names from the xkcd colour survey, then select colours that stayed distinguishable in print. Previously, I’d named the colours after generating the palette, ending up with names like ochre and wine. Those were the closest names I could find, but teachers and students might call them something else. If a teacher says “who has cat? it’s a green one”, everyone needs to know which colour they mean.

Each candidate swatch now starts at the centroid for a colour name in the survey, which collected people’s free-form names for colours. The search selects eight: black, grey, red, brown, green, blue, purple, magenta. Six print at their survey centroid exactly; grey and green move just far enough to stay legible on paper.¹

The cutouts now use the same eight colours, so teachers can use the printed key to call out a colour alongside a word.

Eight is close to the ceiling, too. The obvious ninth is orange, and a printable orange sits ΔE 0.074–0.101 from red once it has been through a CMYK profile — you can have orange or red, not both. Green and turquoise collide the same way, as do magenta and pink.

The key is printed on the brief for both the sheets and the cutouts. Students can use it to narrow their search before reading the tokens.